On the number of variables to use in principal component regression

We assume, withoutlossofgenerality, that isdiagonal. The function m anditsderivativem0 arewell-definedandpositiveatz =0 (andhenceR ( ) is well-definedforall > ).